Аннотации:
© 2019 IOP Publishing Ltd. All rights reserved. The possibilities of using fractal analysis for detecting atypical areas on digital images of feldspar (cracks, kinks, adhesions and structural changes) are investigated. The selection of texture heterogeneities and morphological features is carried out using the field of fractal dimension D, the self-similarity coefficient field K, and the measure field SRGB (triangle areas in the coordinate system of properties). The influences of the displacement step as well as the size of the scanning "window" are estimated. The proposed segmentation of digital images of minerals allows one to eliminate the regions with different ranges of fractal measure values and to describe complex multifractal surfaces of minerals taking into account the spectrum of fractal dimensions, because such fractals, along with geometric features, possess certain statistical properties. The application of the binary transformation of the obtained fractal measure fields to the selected threshold value revealed a sensitivity to a quantitative and qualitative description of the complexity of structural and texture boundaries of the surface. This fact which can be used to predict the results of the ores crushing.
